A processing workload (e.g., for a cloud-based application) may be partitioned among multiple processing nodes in a distributed computing system, such that different processing nodes process different portions of the workload. Different processing nodes in the distributed computing system may have different processing capabilities.